Deconvolution technique is currently available for application in spectral analysis to enhance the resolution of spectra, especially the spectra with multi-longitude modes. Iterative deconvolution and Wiener filtering techniques are discussed. The root-mean-square errors of iterative deconvolution and Wiener filtering with different spectral sampling rates are compared via computer simulation. The results indicate that the resolution of the observed spectrum is improved by using iterative deconvolution and Wiener filtering. The iterative deconvolution works better than Wiener filter at low spectral sampling rate. Experimentally, the spectra of two He-Ne lasers centered at 632.8 nm with single mode and multi-mode are deconvolved. The number of longitude modes, longitude spaces and spectral bandwidths of the He-Ne lasers are obtained by using iterative deconvolution. The deconvolved spectra of He-Ne lasers measured with a low-resolution spectrometer are in agreement with the spectra measured directly with a high-resolution spectrometer.